Adaptive Hybrid AI for Enhancing Decision Reliability and Stability in Urban Decision Support Systems

The dataset presented in this study represents a synthesized and integrated urban data environment designed to support the evaluation of an adaptive hybrid artificial intelligence framework for decision support systems. It combines multi-domain features including urban energy demand, traffic flow dynamics, and environmental conditions, reflecting realistic interactions within smart city infrastructures. The dataset is structured as a time-series with hourly observations spanning from January to March 2024, capturing temporal patterns such as daily cycles, peak demand periods, and traffic congestion behavior.

Energy-related variables include electricity load, peak demand, lagged consumption, and rolling averages, simulating data typically obtained from smart meter systems. Traffic-related features encompass vehicle counts, congestion indices, and flow variability, representing mobility patterns in urban transportation networks. Environmental attributes such as temperature, humidity, and rainfall are incorporated to model external influences on both energy consumption and traffic conditions. Additionally, cross-domain interaction features are engineered to capture the interdependency between subsystems, enabling more realistic modeling of urban dynamics.

The dataset has been preprocessed and normalized to ensure consistency and usability for machine learning and hybrid AI applications. It is suitable for a wide range of research tasks, including time-series forecasting, decision stability analysis, hybrid AI modeling, and smart city analytics. While the dataset is synthetically generated, it is designed to closely mimic real-world urban data characteristics, ensuring reproducibility, transparency, and compliance with open data sharing standards. This makes it particularly appropriate for benchmarking adaptive and explainable AI approaches in urban decision support contexts.